The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes Python scripts (init_skill.py, package_skill.py) intended for local execution to automate the skill development lifecycle. These scripts perform standard file system tasks such as creating project structures and archiving files.\n- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the creation and processing of new skill files based on user-provided metadata and content. It incorporates security measures to handle untrusted input, including ingestion of user-supplied skill names via CLI, boundary markers in SKILL.md, and file system capabilities for directory and archive creation. Risks are mitigated via path traversal checks, symlink rejection, and identifier normalization.\n- [SAFE]: Security best practices are followed throughout the provided scripts, such as the use of yaml.safe_load() for parsing configuration and the enforcement of length and character constraints on user-controlled fields in metadata.

skill-creator